Ice-sheet velocity mapping: A combined interferometric and speckle-tracking approach

Abstract

The first and second RADARSAT Antarctic Mapping Missions (AMM-1 and -2) have now acquired interferometric synthetic aperture radar (SAR) over much of the ice sheet. The RADARSAT 24 day repeat cycle is nearly ideal for measuring slow ice motion (e.g. <100 m a 1), but application of SAR interferometry is limited in faster-moving areas. With a 1 day repeat period, ERS-1/-2 tandem SAR data are much better matched to fast motion, but are not always available. Fortunately, several authors have demonstrated the ability to measure velocity in fast-moving areas by tracking SAR speckle from image to image, which works well even in the absence of visible features. While these estimates have intrinsically lower resolution and poorer accuracy than direct phase measurements, they serve well in areas where there are no data suitable for conventional interferometry. This paper describes algorithms I have developed for merging interferometric and speckle-tracking data from multiple swaths to form a single seamless mosaic of velocity. At each point in the mosaic, all the available data are combined to produce estimates of the velocity and the associated error. This technique is demonstrated using RADARSAT data collected over Lambert Glacier, Antarctica, during AMM-1 and -2.